Using Neural Networks to Model Premium Price Sensitivity of Automobile Insurance Customer

2002 ◽  
pp. 41-54
Author(s):  
Ai Cheo Yeo ◽  
Kate A. Smith ◽  
Robert J. Willis ◽  
Malcolm Brooks
Keyword(s):  
Neural Network ◽  
Risk Groups ◽  
Retention Rates ◽  
Price Sensitivity ◽  
Automobile Insurance ◽  
Network Modelling ◽  
Premium Price ◽  
Policy Information ◽  
Cluster A ◽  
One Year

This paper describes a neural network modelling approach to premium price sensitivity of insurance policy holders. Clustering is used to classify policy holders into homogeneous risk groups. Within each cluster a neural network is then used to predict retention rates given demographic and policy information, including the premium change from one year to the next. It is shown that the prediction results are significantly improved by further dividing each cluster according to premium change. This work is part of a larger data mining framework proposed to determine optimal premium prices in a data-driven manner.

Identifying Key Fraud Indicators in the Automobile Insurance Industry Using SQL Server Analysis Services

2019 ◽  
Vol 64 (2) ◽  
pp. 53-71
Author(s):  
Botond Benedek ◽  
Ede László
Keyword(s):  
Neural Network ◽  
Decision Tree ◽  
Naive Bayes ◽  
Insurance Industry ◽  
Naïve Bayes ◽  
Sql Server ◽  
Categorical Variables ◽  
Automobile Insurance ◽  
Price Determination ◽  
Mining Tool

Abstract Customer segmentation represents a true challenge in the automobile insurance industry, as datasets are large, multidimensional, unbalanced and it also requires a unique price determination based on the risk profile of the customer. Furthermore, the price determination of an insurance policy or the validity of the compensation claim, in most cases must be an instant decision. Therefore, the purpose of this research is to identify an easily usable data mining tool that is capable to identify key automobile insurance fraud indicators, facilitating the segmentation. In addition, the methods used by the tool, should be based primarily on numerical and categorical variables, as there is no well-functioning text mining tool for Central Eastern European languages. Hence, we decided on the SQL Server Analysis Services (SSAS) tool and to compare the performance of the decision tree, neural network and Naïve Bayes methods. The results suggest that decision tree and neural network are more suitable than Naïve Bayes, however the best conclusion can be drawn if we use the decision tree and neural network together.

scholarly journals Association of the Lung Immune Prognostic Index with Immunotherapy Outcomes in Mismatch Repair Deficient Tumors

Cancers ◽  
2021 ◽  
Vol 13 (15) ◽  
pp. 3776
Author(s):  
Edouard Auclin ◽  
Perrine Vuagnat ◽  
Cristina Smolenschi ◽  
Julien Taieb ◽  
Jorge Adeva ◽  
...  
Keyword(s):  
Checkpoint Inhibitors ◽  
Performance Status ◽  
Prognostic Index ◽  
Predictive Marker ◽  
Risk Groups ◽  
Two Factors ◽  
One Year ◽  
Metastatic Sites ◽  
Tumor Types

Background: MSI-H/dMMR is considered the first predictive marker of efficacy for immune checkpoint inhibitors (ICIs). However, around 39% of cases are refractory and additional biomarkers are needed. We explored the prognostic value of pretreatment LIPI in MSI-H/dMMR patients treated with ICIs, including identification of fast-progressors. Methods: A multicenter retrospective study of patients with metastatic MSI-H/dMMR tumors treated with ICIs between April 2014 and May 2019 was performed. LIPI was calculated based on dNLR > 3 and LDH > upper limit of normal. LIPI groups were good (zero factors), intermediate (one factor) and poor (two factors). The primary endpoint was overall survival (OS), including the fast-progressor rate (OS < 3 months). Results: A total of 151 patients were analyzed, mainly female (59%), with median age 64 years, performance status (PS) 0 (42%), and sporadic dMMR status (68%). ICIs were administered as first or second-line for 59%. The most frequent tumor types were gastrointestinal (66%) and gynecologic (22%). LIPI groups were good (47%), intermediate (43%), and poor (10%). The median follow-up was 32 months. One-year OS rates were 81.0%, 67.1%, and 21.4% for good, intermediate, and poor-risk groups (p <0.0001). After adjustment for tumor site, metastatic sites and PS, LIPI remained independently associated with OS (HR, poor-LIPI: 3.50, 95%CI: 1.46–8.40, p = 0.02. Overall, the fast-progressor rate was 16.0%, and 35.7% with poor-LIPI vs. 7.5% in the good-LIPI group (p = 0.02). Conclusions: LIPI identifies dMMR patients who do not benefit from ICI treatment, particularly fast-progressors. LIPI should be included as a stratification factor for future trials.

Sign in / Sign up

Export Citation Format

Share Document